GSA DATA REPOSITORY 2138 Data Repositor Hampel et al., page 1/5 SETUP OF THE FINITE-ELEMENT MODEL The finite-element models were created with the software ABAQUS and consist of a 1-km-thick lithosphere, which is subdivided into a 15-km-thick elastic upper crust, a viscoelastic lower crust (15 km thick), and a viscoelastic lithospheric mantle (Fig. DR1). In our reference model, we use a linear viscosit of 1 2 Pa s for the lower crust. Additional models were computed with viscosities of 1 18, 1 19, and 1 21 Pa s, respectivel. We do not appl a strain-rate dependent viscosit to keep the models computationall feasible. In the upper crust, a 4-km-long fault is embedded (Fig. DR1), which is modeled as a contact interface between the fault footwall and hanging wall. Slip does not etend beond the 4-km-wide fault plane. The fault dip is 6 in the normal fault model and 3 in the thrust fault model. Note that our model fault represents an intra-continental dip-slip fault and is not designed for application to the plate interface in subduction zones. Slip on the model fault is controlled b a Mohr-Coulomb criterion τ = C + µσ n, where τ is the shear stress, C cohesion, µ the friction coefficient and σ n the normal stress. In our models, we use a friction coefficient of µ =.4 and zero cohesion in the model phases during which the fault accumulates slip (eplained below). Each model run consists of a sequence of quasi-static analsis steps and starts with the establishment of isostatic equilibrium. Afterwards, slip on the fault is initiated b etending or shortening of the model at a rate of v = 6 mm/a (Fig. DR1). At this stage, the fault slips continuousl and its slip rate, which we do not prescribe, increases graduall until it reaches a constant value. When the fault has reached a constant, stead-state slip rate (measured at the fault center on the model surface), we lock the fault, i.e. further slip is prohibited, while the far-field shortening or etension continues. We prescribe the length of the locked phase such that the 4-km-long fault eperiences a maimum coseismic displacement of 2 m at the fault center during the net model step, which represents the coseismic phase. A coseismic displacement of 2 m reflects the amount of slip tpical of a M w 7 earthquake on a 4-km-long fault. To simulate the earthquake, we change the boundar condition on the fault from locked to unlocked, which enables sudden slip on the fault plane owing to the strain that was built up during the preceding locked phase. In all models, the coseismic phase is prescribed to be 3 s long and most of the slip occurs during the first half of the coseismic phase. The friction coefficient during the coseismic phase is the same as during the analsis step with continuous fault slip (µ =.4). At the beginning of the subsequent postseismic phase, we lock the fault again while shortening or etension of the model continues. Postseismic afterslip on the fault plane is not considered in our models. g = 9.81 m/s 2 15 km Etension or shortening at v = 3 mm/a 15 km Etension or shortening at v = 3 mm/a Upper crust ρ = 27 kg/m 3 E =.5 1 11 Pa ν =.25-15 km -3 km -1 km Z Y length: 4 km X Upper crust Lower crust Lithospheric mantle P litho = 3 1 9 Pa ρ asth = 32 kg/m 3 Lower crust ρ =29 kg/m 3 E =.5 1 11 Pa ν =.25 η = 1 2, 1 18, 1 19 or 1 21 Pa s Lithospheric mantle ρ = 33 kg/m 3 E = 1.5 1 11 Pa ν =.25 η = 1 23 Pa s Figure DR1. Perspective view of the model lithosphere. The size of the model was chosen such that the results are not affected b the model boundaries. The rheological parameters are densit (ρ), Young's modulus (E), Poisson's ratio (ν) and viscosit (η). In the upper crust, a 4-km-long fault is embedded, whose dip is 6 in the normal fault model and 3 in the thrust fault model. Slip on the fault is initiated b etension or shortening of the model at a total rate of v = 6 mm/a. In our models, we use a friction coefficient of µ =.4 and zero cohesion in the analsis steps with an unlocked fault (see tet for details). Gravit is included as a bod force. Isostas is implemented b adding a lithostatic pressure (P litho ; black arrows) and an elastic foundation (springs), whose propert represents the densit of the asthenosphere (ρ asth ), to the bottom of the model, which is free to move in the vertical and the horizontal directions; model sides in the z-plane are fied in the -direction. Fault
Data Repositor Hampel et al., page 2/5 Figure DR 2 Viscosit of the lower crust 1 19 Pa s A) Normal fault B) Thrust fault 1 a 1 a 2 a 1 a 1 a 2 a 2 km 2 1-1 -2 Velocit in the -direction (mm/a) Figure DR2. Postseismic evolution of the fault-parallel surface velocit field for a viscosit of the lower crust of 1 19 Pa s. A: Normal fault model. B: Thrust fault model. All velocities are averaged over a period of 1 a, e.g., the velocit at 1 a after the earthquake is the average over the time interval from 9 to 1 a. White arrows indicate the direction of the total horizontal movement beond the fault tips where fault-parallel motion eceeds fault-perpendicular motion.
Data Repositor Hampel et al., page 3/5 Figure DR 3 Viscosit of the lower crust 1 21 Pa s A) Normal fault B) 1 a 1 a 2 a Thrust fault 1 a 1 a 2 a 2 km.4.3.2.1 -.1 -.2 -.3 -.4 Velocit in -direction (mm/a) Figure DR3. Postseismic evolution of the fault-parallel surface velocit field for a viscosit of the lower crust of 1 21 Pa s. A: Normal fault model. B: Thrust fault model. All velocities are averaged over a period of 1 a, e.g., the velocit at 1 a after the earthquake is the average over the time interval from 9 to 1 a. Arrows indicate the direction of the total horizontal movement beond the fault tips where fault-parallel motion eceeds fault-perpendicular motion.
Data Repositor Hampel et al., page 4/5 Figure DR 4 A) Normal fault 2 m Total slip on fault plane 4 km B).2.4.6 -.4 -.2.2.4 Total horizontal surface displacement (m) Surface displacement in the -direction (m) Thrust fault -.4 -.2.2.4 Strike-slip displacement 2 m Total slip on fault plane 4 km.2.4.6.8 -.2.2 -.2.2 Total horizontal surface displacement (m) Surface displacement Strike-slip displacement in the -direction (m) C) Normal fault D) Thrust fault 1 a 1 a 2 a 1 a 1 a 2 a Viscosit of the lower crust 1 2 Pa s.3.2.1 -.1 -.2 Velocit in the -direction (mm/a) -.3 2 km.2.1 -.1 -.2 Velocit in the -direction (mm/a) 2 km Figure DR4. Results for a 45 -dipping model fault. (A, B): Coseismic total horizontal surface displacement, fault-parallel surface displacement, and strike-slip displacement on the fault plane in (A) the normal fault model and (B) the thrust fault model with a viscosit of the lower crust of 1 2 Pa s. In the left column, the total horizontal surface displacement is shown b color coding as well as displacement vectors. In the right column, the color code refers to the strike-slip on the fault plane, while the vectors show the total slip on the fault plane (maimum dip slip in the fault center: 2 m). (C, D): Postseismic evolution of the fault-parallel surface velocit field for a viscosit of the lower crust of 1 2 Pa s. C: Normal fault model. D: Thrust fault model. All velocities are averaged over a period of 1 a, e.g., the velocit at 1 a after the earthquake is the average over the time interval from 9 to 1 a.white arrows indicate the direction of the total horizontal movement beond the fault tips where fault-parallel motion eceeds fault-perpendicular motion.
Data Repositor Hampel et al., page 5/5 Figure DR 5 Oblique thrust fault 4 km.5 1. Total horizontal surface displacement (m) -.1.1.2.3.4 -.2 -.1.1.2.3 Strike-slip displacement Surface displacement in the -direction (m) Figure DR5. Coseismic total horizontal surface displacement, surface displacement in the -direction, and strike-slip displacement on the fault plane in an oblique thrust fault model with a viscosit of the lower crust of 1 2 Pa s. The angle between fault strike and shortening direction is 5. The dip-slip displacement on the fault plane is 2 m at the model surface. The total horizontal surface displacement is shown b color coding as well as displacement vectors.